When something is not working in prod, the first thing I do is check basic operational stuff with the underlying deployment. Are all the pods still running? Have there been any restarts? If there is some DNS/network error how can I spin up a pod in the cluster to check on various things?

And how much less downtime would you have if domain experts were doing that part?